That's a shame, because Pandas has a really quirky/legacy-burdened API and Polars is super clean by comparison. As someone who had Spark and Pandas experience before switching, Polars felt like Pyspark without the added mental overhead of needing you to think about multi-worker-node parallelism
replyIt's just muscle memor, I've been using Pandas daily for over a decade, but I'll likely eventually switch over to Polars.
replyPart of the problem, as I said, is that I'm just spoiled by DuckDB when performance matters.